Sverker Eriksson

Oh, I didn't think about the putting-things-last optimization in copy_struct. Hm.... _I'm sorry if I persist skeptical of your quite impressive `copy_struct_in_order_x()`, but as a maintainer I'm a bit stickler...

The idea of this PR was to optimize process heap memory consumption at the expense of CPU load caused by more frequent GCs when a process **creates lots of live...

> I'm happy to try a patch out if there's still interest in this. This branch is currently based on OTP-24.3, what do you need? You can always try merge...

To be continued. Get the map comparison correct as documented.

Forced pushed a working version rebased on a fresh master.

Squashed into one commit to soon be merged to master.

Merged to master. Thanks @josevalim for the initial idea, commit and benchmarks. To be continued in #6718 to restore some order.

@RobinMorisset ping

Fixed bug causing failed debug assert in `ets_SUITE:lookup_element_mult`.

@RobinMorisset I've fixed some problems with `ets:update_element` and the empty tuple (which is always a literal). Now I'm satisfied and will let it run at least one night before squash...